On 20 Oct 2020 Atlassian announced the future discontinuation of its Server products in favour of the Cloud equivalents.

The response has been overwhelmingly negative on Atlassian's forums [1, 2] and publicly [3, 4, 5, 6]. Objections are in roughly 3 categories:

• Some organizations unable to move to Cloud due to:
  • security / privacy concerns
  • regulatory concerns (e.g. server hosting jurisdiction).
  • existing deep integrations with internal apps/data
• Dislike of Atlassian's Cloud products, for being:
  • 'dumbed down'
  • slow
  • inflexible (fewer, weaker add-ons)
  • unstable (regular UI changes; third-party app stability)
• Price increases
  • Cloud is relatively unaffordable. Cost increasing 3x to 5x vs. Server renewals.

I too am deeply disappointed. I have long believed the Server products are simply better than Cloud equivalents. Red Radish Consulting's schtick has always been helping companies getting the most out of Server products; so what happens now they're going away? Atlassian are killing the better product to force people onto their higher revenue product, to the detriment of everyone.

Isn't this what happens when good companies take VC money: they sell out?

It would be so easy to jump on the hate and cynicism bandwagon. But let's consider:

• No company has an obligation to support a particular piece of software forever. We can disagree with Atlassian's choice of direction, but ultimately it's their software. Let's gracefully accept it and move on. Any stronger emotion is probably the result of a misplaced sense of entitlement.
• We were warned. Atlassian have been very vocal for years that they regard Cloud as the future. Development in Server has been stagnant for the last 10 years.
• We should assume good faith. I truly believe Mike and Scott want to do what's best for their customers. This is not a cynical cash grab under VC pressure. I was in the room when Atlassian's Values were hashed out: they truly believe in "don't f**k the customer". Mike and Scott believe that:
  • Cloud is a fundamentally better way of delivering software
  • By focusing on Cloud, Atlassian can build better Cloud products
  • Some short term pain now will result in more, happier customers in the medium to long term.

There are counterarguments to all these points. 3 years' notice is not particularly generous for a 17 year old product. While it's true we were warned, most people expected a dignified retirement for Server rather than being taken out the back and shot.

As for Atlassian's strategy, it hinges on the theory that killing Server will somehow make Cloud a better product. This doesn't fly with me for two reasons.

Firstly, Atlassian have had 10 years and all the money they could want to make Cloud good. They've frittered it all away chasing UX fads and creating dumbed-down 'next-gen' interfaces. Shutting down Server will let Atlassian "focus on Cloud" – but will having more cooks in the kitchen really help? Does anyone remember The Mythical Man-Month?

Secondly, and more fundamentally: Cloud may be a better software delivery model, but it delivers a standardized, one-size-fits-all product which for a significant minority is a worse overall experience, or cannot be used at all. It's like McDonalds vs. home cooking: eating out is certainly more convenient, but you can certainly eat better if you know how to cook, and some people with special dietary needs can't eat McDonalds food at all. The world would not be better off if the only food was take-away; likewise the industry trend to SaaS is not making user lives better.

Also, eating out every day gets really expensive. I'll tackle the issue of pricing in a separate post.

As painful as it is, perhaps Atlassian's announcement is for the best. Server development has been stagnant for many years. The rising frustration of customers can be seen on the open feature requests on https://jira.atlassian.com. Saying "we're not developing this any more" is more honest than continuing to take people's money without fulfilling the implicit promise to keep developing features.

Where to go from here? I don't rightly know yet. This is a perfect opportunity to evaluate the options, both SaaS, open source and hybrid ("open core"). I started a forum for Atlassian Server product refugees to evaluating options together, over at:

https://www.goodbyeserver.org/

Please join us and say hi, or at least subscribe to the newsletter.